Multiplex Detection of Pancreatic Cancer
In collaboration with N. Banaei, J. Moshfegh, Y. Sun, B. Kim
Multiplex Detection of Pancreatic Cancer Biomarkers
In this work which is an extension of this paper authored by my collaborators, we are using novel experimental and analytical techniques to maximize the accuracy of early diagnosis of various cancer types. The data is collected by sampling the bloods of the test subjects and using gold nanopartickes and specific substrates to detect specific biomarkers. Raman spectroscopy is then used to read the density of certain biomarkers in the blood sample. Given this data for various biomarkers, we have a supervised learning task to classify the blood samples as cancerous and non-cancerous.
My Research Contribution
I am using machine learning techniques to enhance the analytical methods used to analyze SERS based Raman spectral data. So far, we have achieved a very high accuracy in predicting Pancreatic and Ovarian cancer compared to the state of the art analytical methods by transforming the Raman spectral data. In these method, I am using this transformations along with methods like k-nearest naighbors and gradient boosting trees to gain higher accuracies.
